Following the reception was the PGP keysigning. Building up the ASF web of trust is important, so I'm glad that we did an official event this year. In the past, this has always been done informally at the hackathon. Unfortunately, some of the people who didn't sign up, and a bunch more people showed up unprepared. Next year, we need to have PGP tutorial session before the actual keysigning.

After the Atom BOF, which was relatively uneventful (lots of discussion), I went to dinner with Ben Hyde, Ben Laurie, and Cliff Skolnick, all httpd oldtimers. We had an interesting programming language discussion (among others) that covered APL, SL5, E, Snobol, and Lisp.

By the time we finished dinner, it was kind of late, but I still had some e-mail, etc to do, so I headed over to the online lounge to take care of that. A group of ASF'ers came back from their dinners, and I ended up staying in the lounge until about 2:30AM. Brian Behenldorf and Manoj Kasichainula educated me on BitTorrent and Tivo Hacking. Dirk-Willem van Gulik spent some time telling PowerBook hints -- I need more contact with the hardcore FreeBSD users who are on OS X.

Dirk is involved in a bunch of stuff. I've written before about the fiber optic backbone that is being installed on Bainbridge Island, with the potential of fiber drops to people's homes. The last mile process is moving more slowly than I would like, and Dirk shared some experiences that they've had with fiber in Leiden. More interesting is that Dirk's been involved with WirelessLeiden, which has installed a wireless network that provides coverage for a city of 80,000 -- 4 times larger than the city of Bainbridge Island. The most important thing that I learned from Dirk is that trees are big problem for wireless networks. Since we have a lot of trees on the island, I'm pretty pessimistic that a wireless solution is going to work effectively for Bainbridge Island.

After that, conversation turned to Chandler, and I discovered that Dirk's company is working on stuff in the semi-structured repository space. Now I have a few reasons to keep in better touch with Dirk...

The day was dominated by the fact that I did two sessions back to back -- well except for the lunch break in between. The first session, "XML at the ASF" is a new presentation that is an overview of all the XML related projects at the ASF. It's a brand new presentation based on the contents of my book, (which unfortunately didn't escape from the printer in time to be ready for ApacheCon). Doing a presentation the first time is always an adventure, so I was happy to end on time, and not get hecked to much. There was a glitch with my Thinkpad not finding the external video output, which resulted in the obligatory reboot. It seemed to go well, and people told me that it was good. Overview talks are hard, because you are basically saying not very much about a bunch of things, which just isn't my style -- I like to have the details.

After my talk, a bunch of people came up to test their laptops (having seen my troubles), and I got to meet Mark Pilgrim. I spent lunch talking with Mark, Joe Gregorio, and Stefano.

The second talk "Everything you ever wanted to know about XML parsing", is a standard tutorial on how to use Xerces. This presentation has appeared at the last 3 ApacheCon's, with different lengths each time, according to the duration of session. This year the sessions are 50 minutes long, which makes it hard to put a lot of detail into the talk. I had just finished the section on SAX, when Lars Eilebrecht, the session chair, held up the 20 minutes left sign. I started talking faster -- so I ended up rushing to finish. Someone later told me that I had already been talking fast, and that made it even faster. At least he told me that he got a lot of useful information out of the talk.

After I got down from the podium, David Bau, one of the XMLBeans committers got up and did a talk on XMLBeans. His presentation was a convincing demonstration of the utility of XMLBeans. In fact, it was so convicing that two people gave David a standing ovation! I was gratified to see that most of the conversation in the IRC channel was highly laudatory. Afterward, I know that Andy Oliver and Brian McAllister were convinced that XMLBeans was it.

Next up was the Mark and Joe show. Mark delivered a talk on the state of the Atom API (not the syndication format). His talk took the form of a history lesson and technical analysis of the various weblog APIs, followed by an explanation of the Atom API. This was really helpful to me, because I haven't been able to keep up with the progress on Atom. So now I know what the state of the API is.

Joe followed Mark and talked about the way that they've extended HTTP's authentication model to enable Atom to do authentication in the typical (hosted) environments where Atom API implementations are likely to run. He did this by explaining the various revs of the API and the mistakes that caused revisions to be made. It must have been a slightly intimidating experience for Joe, because Roy Fielding ws sitting in the middle of the room. At one point Roy interrupted Joe to point out a way that they were mis-using "X-" headers. But at the end, Roy basically told Joe and Mark that they were doing it the right way. Now that's what I call validation, and its the kind of wonderful face to face thing that happens at ApacheCon.